• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

That pesky 502 bad gateway nginx error again :-(

theWoosh

Basic Pleskian
Hi
I have an ongoing problem that is:

When I access plesk control panel as admin user it works fine, if I go to the control panel for particular domains, some of them work fine and for some I get the 502 error. After that I cannot access plesk at all - just get the same error until I start a fresh session.

The errors in var/log/sw-cp-server/error_log are ones like:
[error] 20661#0: *23 recv() failed (104: Connection reset by peer) while reading response header from upstream, client: 10.255.250.9, server: , request: "GET /smb/ HTTP/1.1", upstream: "fastcgi://unix:/var/run/sw-engine.sock:", host: "blah.com:8443", referrer: "https://blah.com:8443/admin/subscription/overview/id/35"

I have been unable to find anything online that is useful to diagnose this very specific problem. I have long ago modified /etc/sw-cp-server/config to read:
fastcgi_buffers 128 2048k;
fastcgi_buffer_size 2048k;

Can anyone suggest a way to find out what is going on and fix this problem?
I don't have any support for this as the company who provide hosting don't have any knowledge of PLesk, so I'm on my own... :-(

Many Thanks,
theWoosh
 
sorry should say I'm on Ubuntu 14.0.4 LTS, Apache 2.4.7, nginx 1.6.0-1, mysql 5.6.29 and Plesk 12.0.18
 
Did you upgrade Plesk? It was working fine and now it broke randomly?
Are all service are running? nginx/apache/psa/etc?
 
Not sure when it got broke - could have been a while back as the domains I am trying to access are older ones that are little used. This plesk install is a migration from an earlier version on a different server BTW, so may of the domains have been migrated (but there doesn't seem to be a patter as to which ones produce the error and which ones don't). Everything else on the server is working fine Plesk works fine as long as you don't try to access control panels for the affected domains, all of the domains are serving web pages including the affected ones....
 
Upgraded to latest version of Plesk 12.0.18 (#77), but that didn't help.
I tried running /usr/local/psa/bootstrapper/pp12.0.18-bootstrapper/bootstrapper.sh repair just to see if this helps, but this had problems notably:
WARNING! Some problems are found during fix credentials for APS controller database(see log file...
...
Regenerating web servers' configuration files...Execution failed....
(log shows httpdmng --reconfigure-domains then a list of all domains)
...
Details: sw-engine: /usr/include/libboost-plesk-1.54.0/boost/variant/detail/visitation_impl.hpp:182: typename Visitor::result_type boost::detail::variant::visitation_impl_invoke(int, Visitor&, VoidPtrCV, boost::detail::variant::apply_visitor_unrolled*, NBF, long int) [with Visitor = boost::detail::variant::invoke_visitor<returner>; VoidPtrCV = void*; NBF = boost::variant<long int, bool, std::basic_string<char>, boost::gregorian::date>::has_fallback_type_; typename Visitor::result_type = void]: Assertion `false' failed.
...
Trying to reconfigure web-server configurations skipping broken domains... Execution failed.
(logs show this was for just one domain)
Details: sw-engine: /usr/include/libboost-plesk-1.54.0/boost/variant/detail/visitation_impl.hpp:182: typename Visitor::result_type boost::detail::variant::visitation_impl_invoke(int, Visitor&, VoidPtrCV, boost::detail::variant::apply_visitor_unrolled*, NBF, long int) [with Visitor = boost::detail::variant::invoke_visitor<returner>; VoidPtrCV = void*; NBF = boost::variant<long int, bool, std::basic_string<char>, boost::gregorian::date>::has_fallback_type_; typename Visitor::result_type = void]: Assertion `false' failed.

WARNING!
Some problems are found during reconfigure web-server configurations skipping broken domains(see log file
 
Back
Top